A primer for the easiest job around (for those who don’t do it)
I tell everyone that product management is common sense. Yet you probably need this guide to do it. Until you can reconcile this, you have much to learn.
Step 1: Gird thyself
Product management (PM) can be a wonderful job …at times …usually in startups. In big companies it’s often thankless. More than any other job, your success will depend on situational factors you cannot control. And also more than any other function, no one cares how much effort you put in (unless you’re obviously slacking off); results are the only thing. Of course, this also means that complete idiot PMs will get promoted because they were in the right place at the right time. Developers and data scientists will see their incompetence and generalize across everyone who does product.
Oh yeah, and your job is the only one where everyone including the janitor believes he can do it. Not only that, many of those who express this view believe that they are so superhuman (and you so dispensible) that they can do their job and yours — at the same time!
I have been at multiple companies that already had product managers on the payroll but where the value of product management was hotly debated. I’m not necessarily opposed to this, but can you imagine any other job type where people advocate that the folks doing this job — who they sit across and work with — just not, umm, be around?
Welcome to product management.
Step 2: The secret is there’s no secret
One day AI will do product management, but that’s not saying much. One day AI will do everyone’s job. (Meet AI Tilda.)
PM is a game that’s easy to start, hard to master.
I’ve had new PMs or data scientists ask me for templates, tools, skills, frameworks, training courses. Yes, these things exist. But they’re not the thing.
The truth is that great PMs, just like other great professionals, excel at a few areas of their field — not all of them. Meanwhile, many of these areas are not codified or learnable in the same way that computer science or data science are. The learnings are distributed all across the internet — or, more likely, ingrained in a PM over years of hits, misses, contemplation, and mentorship.
More than anything, the PM’s value consists of leadership. The biggest components of product leadership are ownership, judgment, customer obsession, creativity, bias for action, and strategy. There’s no textbook for these. Yes, they’re “soft” skills. Yes, they’re critically important. Yes, excellence at these is very rare.
And while anything can be learned, certain types of people have more raw material for these than others in my experience.
So if you think you’ll be able to do the job well once you find that decoder ring, think again.
Step 3: Results are all that matter now
Engineering velocity doesn’t matter. Code cleanliness doesn’t matter. Tech debt is less important than before you became a PM. Models going live don’t matter. Having fun exploring the data doesn’t matter. The only thing that matters is results.
Step 4: Results = Value / Time
Value is net of effort. Time is calendar time.
Sometimes value is hard to nail down, but it’s critical that you at least try to measure it with as “high” a business metric as you can. “High” in this case means closest thing to NPV possible, even if sometimes it’s only realistic to get to a much lower-level measure.
Step 5: Have a point of view
You don’t get to just do what other people ask for anymore. (You ideally weren’t just doing that anyway, but now it’s time to double down.)
You are not an account manager for an agency or “dev shop.”
This is especially important when your customers are internal. The fact that a VP asked for X or is happy with Y is less important than you delivering value to the business’ end customers, shareholders, and other key stakeholders if applicable.
Step 6: You own it
You can’t blame the customer for not being clear. You can’t blame designers. You can’t blame leadership. You can’t blame the dev team.
Product managers own stuff. That doesn’t mean they are better than anyone else. It’s merely the role they play. And let’s be clear, leadership is the biggest thing product managers add.
If they don’t own the “what,” major decisions, or the outcome itself, then they’re not really doing product.
You own it now. So act like it.
Step 7: Report up and out
If you own it, you seek input, convey updates, and escalate risks proactively. No one is doing this for you. No, it’s not supposed to be fun or interesting.
Step 8: Paint a vision, and make a business case
You’re not some ronin looking for work for your data science agency. You are in the business of strategy, research, ideas, and building full solutions. This is not only more complex a goal than a simple model or piece of software, it is also considerably more risky and resource intensive.
If all of this sounds exhausting to you, then don’t play product manager. Real product managers live for these things.
Don’t pick the first idea that a customer relates to you, as there are numerous others you could work on. Evaluate the tradeoffs.
And get specific. It’s not enough to say you’re working on your customer’s biggest priority. You need to say how you’re going to do that — and why your solution is the most amazing one. How much value do you estimate you’ll create? Paint a picture of how you and the team will get there. What will your MVP be and and why?
Eventually you’ll need to specify clear and compelling business requirements to the team, but don’t put the cart before the horse. Nail the high level first. If all of this sounds exhausting to you, then don’t play product manager. Real product managers live for these things.
Then get input and approval. These should be obvious. But in the absence of PM I’ve seen technical folks overlook this and operate in a silo. Implicitly they are assuming complete autonomy, though obviously no one in a modern company would ever claim that out loud.
Step 9: Track it like you mean it
Although technical folks and scrum masters often take on the heavy lifting of tracking implementation work, teams without PMs are often sloppy, too high level, or don’t track much at all. You’re the product manager now, so fix it.
Step 10: Do the dirty work
Product managers do “shit work” all the time. Hopefully they don’t complain about it, but it’s the truth.
PMs don’t get to say “not my job.” So if no one is doing a particular function, it’s your job now (yay!)
Step 11: Declare that you’re the product manager
Experienced tech folks know that someone is doing product’s job even if there’s no PM. If you are stepping up, then say so with clarity. Find someone to backfill your new-old responsibilities. Tell your customers, your boss, and your team. Put it in your title.
This will eliminate confusion with other team members and customers. Perhaps more importantly, it will prevent you from deluding yourself that your job is really something different and that it’s implicitly okay to half-ass product.
You can’t have your cake and eat it, too. If you’re doing product, just do it.
 Customer obsession, bias for action, and ownership are Amazon leadership principles. Strategy encapsulates some others but is not an Amazon leadership principle per se. The author spent four years at Amazon.
Sri hosts The Warrior Poet podcast, a show on the philosophy of leadership based on his experience in the SEAL Teams, at Harvard Business School, on Wall Street, and in tech. Shows every Monday. Follow him on Instagram @sri_the_warrior_poet.